**** BEGIN LOGGING AT Fri Nov 22 02:59:57 2019 Nov 22 07:46:57 Morning Nov 22 08:13:12 Morning! Nov 22 08:21:14 Tofe: Did the build succeed now? Nov 22 08:33:01 webruntime is built, yes, but as I did some changes in qtwebengine it also has to be rebuilt :p Nov 22 08:33:16 should be ready soon though Nov 22 08:45:37 Ah ok Nov 22 08:49:09 Chromium seems to build quicker v.s. QtWebengine here, or at least it seems that way Nov 22 09:11:37 build finished, I'll test that this evening Nov 22 09:15:34 :) Nov 22 13:40:47 https://www.phoronix.com/scan.php?page=news_item&px=webOS-OSE-2.1-Released Nov 22 14:34:51 Tofe: the scaling fix from you seems to work fine, any objection merging 5.13 to zeus branch? Nov 22 14:35:43 unstable is building last few rpi images with it Nov 22 14:40:56 JaMa: I've observed a little regression regarding power menu, where the height is incorrect; but we can fix that later, I guess it's also an issue with the scaling of the height Nov 22 14:42:20 is there a way to trigger power menu in emulator? Nov 22 14:42:47 I don't think so, or maybe by starting the app itself Nov 22 14:43:23 maybe via a ls2 call Nov 22 14:49:23 JaMa: "palm://com.palm.display/com/palm/display/powerKeyPressed" with "{\"showDialog\":true}" Nov 22 14:58:58 Or for easy C&P: luna-send -n 1 palm://com.palm.display/com/palm/display/powerKeyPressed '{"showDialog":true}' Nov 22 14:59:35 perfect, thanks both Nov 22 15:01:28 well, almost there {"returnValue":false,"errorCode":-1,"errorText":"Unknown method \"powerKeyPressed\" for category \"/com/palm/display\""} Nov 22 15:06:10 JaMa: Let me check Nov 22 15:06:39 https://github.com/webOS-ports/luna-sysmgr/blob/webOS-ports/master/Src/base/DisplayManager.cpp#L82 ? Nov 22 15:06:41 And? luna-send -n 1 palm://com.palm.bus/signal/addmatch '{"category": "/com/palm/display", "method": "powerKeyPressed"}' Nov 22 15:06:57 I have my Qemu with broken WAM, so it won't show me anything :P Nov 22 15:07:41 getting closer :) {"returnValue":true}, but nothing is shown Nov 22 15:08:05 addmatch will just listen for events Nov 22 15:09:40 Tofe: OK let me grep around 3.0.5 Doctor ;) Nov 22 15:09:44 what about palm://org.webosports.luna/com/palm/display/powerKeyPressed Nov 22 15:09:56 mmh no, looks wrong Nov 22 15:10:17 root@qemux86-64:~# luna-send -n 1 palm://org.webosports.luna/com/palm/display/powerKeyPressed '{"showDialog":true}' Nov 22 15:10:20 {"returnValue":false,"errorCode":-1,"errorText":"Unknown method \"powerKeyPressed\" for category \"/com/palm/display\""} Nov 22 15:10:34 maybe it isn't enabled in emulator? Nov 22 15:11:56 we had that discussion before :D http://logs.nslu2-linux.org/livelogs/webos-ports/webos-ports.20151027.txt Nov 22 15:12:56 mmh but no answer there, it seems Nov 22 15:15:40 Tofe: Seems there's no direct luna-send call probably: https://github.com/webOS-ports/luna-systemui/blob/webOS-ports/master/data/PowerdService.js#L197 Nov 22 15:26:51 Herrie: well this method is activated when it sees a call on the "powerKeyPressed" method on the bus, if I understand correctly Nov 22 15:31:27 Yes Nov 22 15:31:37 We could possibly map a key to trigger it on VBOX Nov 22 15:31:44 To simulate the power button Nov 22 15:39:41 * JaMa didn't expect to open this discussion just to see the Qt height issue Nov 22 15:55:59 :) Nov 22 15:56:13 JaMa: you should also see the issue with dashboards, I guess Nov 22 15:56:16 I didn't test it Nov 22 15:56:39 But it's also possible that the scaling is 1.0 on qemu :) Nov 22 16:09:12 JaMa: Dashboards you can easily test in Testr App Nov 22 16:30:46 Testr I was testing before and it looked ok Nov 22 16:53:00 Could just be the systemui that needs a little tweaking in HTML/CSS Nov 22 16:53:09 The Power Menu was always a bit of a headache Nov 22 18:12:09 Herrie: what do you start for wam ? Nov 22 18:19:50 Tofe: export XDG_RUNTIME_DIR and then the .sh that's in the systemd service file for wam Nov 22 18:20:25 I just found it Nov 22 18:20:30 but the output is a mess Nov 22 18:20:52 https://bpaste.net/show/CXUQ6 Nov 22 18:21:44 in the .sh, I have "mkdir -p ${WAM_DATA_PATH}" Nov 22 18:21:53 and chown -R ${WAM_UID}:${WAM_GID} ${WAM_DATA_PATH} Nov 22 18:22:04 where nothing seems to be defined Nov 22 18:22:21 -- dinner time -- Nov 22 18:22:32 Hmmz I had pwd before as well, I added something to basepwd I think. Pretty sure I pushed that Nov 22 18:23:10 Well WAM_DATA_PATH it also pulls in the .sh somehow if I recall correctly? Nov 22 18:23:19 I haven't seen this exact error before Nov 22 18:23:29 Nothing busybox for sure Nov 22 18:23:51 Or WAM_DATA_PATH should be an env variable? Nov 22 18:24:50 not sure, let me see Nov 22 18:25:31 UID/GID are in the .sh, that's fine Nov 22 18:25:47 ah: WAM_DATA_PATH=$(grep ${WAM_UID} /etc/passwd | cut -f6 -d":") Nov 22 18:26:04 let me add the "wam" user Nov 22 18:28:46 Herrie: could you show me the corresponding line in OSE qemu's /etc/passwd please ? Nov 22 18:30:21 or I could deduce it from the recipe, where EXTRA_QMAKEVARS_PRE += "DEFINES+=WAM_DATA_DIR=\"\"${webos_cryptofsdir}/.webappmanager/\"\"" Nov 22 18:31:45 I thought I pushed the basepwd stuff to meta WIP? Nov 22 18:31:50 Meta-wop Nov 22 18:32:11 Or maybe I did it in meta-smartphone and still need to push it Nov 22 18:32:16 Let me check Nov 22 18:32:19 ok :) Nov 22 18:35:00 Seems I did something in meta-smartphone Nov 22 18:35:03 Let me push those changes Nov 22 18:35:30 Gimme a few minutes ;) Nov 22 18:35:38 no pb Nov 22 18:45:21 Tofe: https://github.com/Herrie82/meta-smartphone/commit/24556ced04755d67a5ffb74d7d40fe0bbf27d095 Nov 22 18:45:29 Needed to add Zeus branch to my fork, so took a little longer ;) Nov 22 18:46:54 well this should go to meta-wop somehow Nov 22 18:47:07 but that'll be enough for me to fix my issue here Nov 22 18:49:29 Herrie: you know what ? Nov 22 18:49:33 it doesn't crash here :p Nov 22 18:49:55 well I do get the FATAL error too Nov 22 18:50:04 but the process doesn't end Nov 22 18:50:27 Ok Nov 22 18:50:33 Well that's good Nov 22 18:51:26 now, let's crush this display error Nov 22 18:53:27 You have the same error as I have? Nov 22 19:04:59 yes Nov 22 19:05:10 but now I can use gdb ;) Nov 22 19:18:16 connect(46, {sa_family=AF_UNIX, sun_path="/tmp/luna-session/wayland-0"}, 30) = -1 EACCES (Permission denied) Nov 22 19:19:32 ah, ok, on OSE the compositor has the guid "compositor" and creates the wayland socket; and "wam" guid is in that group Nov 22 19:23:32 Ah ok, so it seems to be mainly permissions then? Nov 22 19:23:44 probably Nov 22 19:23:49 I'll try to force my way Nov 22 19:24:51 mmmh "listener function for opcode 3 of wl_output is NULL" Nov 22 19:28:29 That's like Chinese to me. Could be their Wayland extensions somehow? Nov 22 19:29:05 it's when chromium's ozone wayland implementation tries to guess the display geometry in OzoneWaylandScreen::LookAheadOutputGeometry Nov 22 19:29:26 I'm also trying to decipher the message :p Nov 22 19:30:06 https://bpaste.net/show/X5ODO that's where we are now Nov 22 19:37:57 I saw AGL passing a resolution in Nov 22 19:38:12 Check Skype LuneOS group for link to their recipe Nov 22 19:39:13 ok Nov 22 19:39:29 I think in the Chromium BB or .inc Nov 22 19:39:55 well it's not *that* bad: wl_output@7.mode(0, 1080, 1920, 60000) Nov 22 19:40:04 the client does receive something in Nov 22 19:47:40 ok, if I'm correct the issue we have is that: 1. webruntime is compiled with luneos's wayland, which is very recent; 2. therefore, is declares that it uses the very latest version of the protocol; 3. but in one of the callbacks, it doesn't fill up all the needed callback for the most recent protocol version Nov 22 19:48:02 Adding an empty function for the "scale" opcode should do it Nov 22 19:50:05 however it could be that it's only the first of our surprises... Nov 22 19:53:14 webruntime rebuilds quite fast ! that's nice :) Nov 22 19:56:46 The safest bet might be to use webruntime's shipped wayland version though Nov 22 19:59:17 Tofe: They have a flag to use upstream wayland Nov 22 19:59:20 See AGL Nov 22 20:00:41 well I don't want to rebuild the whole thing... first I'll just add this little callback, and if I run into another issue then I'll switch to the internal wayland Nov 22 20:01:24 but even OSE's WAM might have this kind of issue if they aren't careful enough Nov 22 20:04:53 Herrie: well... it doesn't crash now! Nov 22 20:05:04 so, what now? Nov 22 20:39:20 Tofe: Well does an app launch? Nov 22 20:54:07 nope Nov 22 20:54:18 but there are multiple permission issues Nov 22 21:10:35 Such as? Nov 22 21:10:47 Some might be easy to fix Nov 22 21:16:02 Herrie: https://bpaste.net/show/FFXFU is what I get in the logs Nov 22 21:18:41 I saw the Bootmanager one. Easy to fix, will make sure that FirstUse gets launched Nov 22 21:19:01 Others are due to non existing API calls Nov 22 21:20:30 And in general they shouldn't be critical Nov 22 21:27:15 I'll investigate more tomorrow Nov 22 21:30:39 I can push the 2 permission fixes I have in my notes Nov 22 21:36:40 Tofe: https://github.com/webOS-ports/bootd/commit/4b454dd8394852db751f7f2de5936480e3e35b1a Nov 22 21:36:48 That's the one that will boot to FirstUse Nov 22 21:36:58 thanks Nov 22 21:37:01 Other one is for pmloglib and rdxd, not very important Nov 22 21:37:07 ok Nov 22 21:38:00 sregisterForAppEvents is something we added in our luna-webappmanager it seems, didn't exist in OWO Nov 22 21:38:53 I think the "listApps" issue might be the main one Nov 22 21:39:24 Well that is also in our current images I think Nov 22 21:39:34 oh? ok Nov 22 21:39:34 It's not WAM specific from what I recall Nov 22 21:39:41 But I might be wrong Nov 22 21:40:52 It's probably because we have luna-sysmgr still and not SAM Nov 22 21:41:59 Change com.webos.applicationManager in https://github.com/webosose/wam/blob/2e59f64c6565c942284469b22b61d838e7c2d3fe/src/webos/WebAppManagerServiceLuna.cpp#L538 into com.palm.applicationManager and we should be good there Nov 22 21:42:21 Though we might get other permission issues then ;) Nov 22 21:44:59 But those should be easy to fix **** BEGIN LOGGING AT Fri Nov 22 22:22:56 2019 **** ENDING LOGGING AT Sat Nov 23 02:59:57 2019